Optimization of minoxidil microemulsions using fractional factorial design approach.

http://informahealthcare.com/phd ISSN: 1083-7450 (print), 1097-9867 (electronic) Pharm Dev Technol, Early Online: 1–12 ! 2014 Informa Healthcare USA, Inc. DOI: 10.3109/10837450.2014.971375

RESEARCH ARTICLE

Pharmaceutical Development and Technology Downloaded from informahealthcare.com by University of Auckland on 10/16/14 For personal use only.

Optimization of minoxidil microemulsions using fractional factorial design approach Napaphak Jaipakdee1, Ekapol Limpongsa1,2, and Thaned Pongjanyakul1 1

Division of Pharmaceutical Technology, Faculty of Pharmaceutical Sciences, Khon Kaen University, Khon Kaen, Thailand and 2Center for Research and Development of Herbal Health Products, Khon Kaen University, Khon Kaen, Thailand Abstract

Keywords

The objective of this study was to apply fractional factorial and multi-response optimization designs using desirability function approach for developing topical microemulsions. Minoxidil (MX) was used as a model drug. Limonene was used as an oil phase. Based on solubility, Tween 20 and caprylocaproyl polyoxyl-8 glycerides were selected as surfactants, propylene glycol and ethanol were selected as co-solvent in aqueous phase. Experiments were performed according to a two-level fractional factorial design to evaluate the effects of independent variables: Tween 20 concentration in surfactant system (X1), surfactant concentration (X2), ethanol concentration in co-solvent system (X3), limonene concentration (X4) on MX solubility (Y1), permeation flux (Y2), lag time (Y3), deposition (Y4) of MX microemulsions. It was found that Y1 increased with increasing X3 and decreasing X2, X4; whereas Y2 increased with decreasing X1, X2 and increasing X3. While Y3 was not affected by these variables, Y4 increased with decreasing X1, X2. Three regression equations were obtained and calculated for predicted values of responses Y1, Y2 and Y4. The predicted values matched experimental values reasonably well with high determination coefficient. By using optimal desirability function, optimized microemulsion demonstrating the highest MX solubility, permeation flux and skin deposition was confirmed as low level of X1, X2 and X4 but high level of X3.

Desirability function, fractional factorial design, microemulsion, minoxidil, optimization

Introduction Microemulsions represent pharmaceutically versatile formulations for numbers of applications including drug delivery to and through the skin. A microemulsion is defined as a transparent, single, optically isotropic and thermodynamic stable colloidal dispersion of water, oil and surfactant, in many cases, in combination with co-surfactant1. Microemulsions provide the possible requirements of a liquid system including easy formation, low viscosity with Newtonian behavior, high solubilization capacity and small droplet size, which typically less than 150 nm. Microemulsions are dynamic systems in which the interface is continuously and spontaneously fluctuating. Depend on the type of surfactant and content of the oil used, three types of microemulsions available include oil dispersed in water (O/W), water dispersed in oil (W/O) and bicontinuous microemulsions1–3. Microemulsion formulations showed significant permeation enhancement effect for both lipophilic and hydrophilic drugs4–9. The drug delivery potential of microemulsion depends on the constituents of vehicle and, a significant extent, the composition/ internal phase structure which may retard drug diffusion in the vehicles. Zhang and Michniak-Kohn8 reported that the permeation flux and the cumulative permeation amount after 24 h related to microemulsions microstructures. The permeation flux of

Address for correspondence: Napaphak Jaipakdee, Division of Pharmaceutical Technology, Faculty of Pharmaceutical Sciences, Khon Kaen University, Khon Kaen 40002, Thailand. Tel: +66 81 9749228; + 66 43 362092. Fax: +66 43 362092. E-mail: [email protected]

History Received 29 June 2014 Revised 24 September 2014 Accepted 24 September 2014 Published online 15 October 2014

hydrophobic and hydrophilic model drugs, ketoprofen, lidocaine, and caffeine from microemulsions were in an increasing order of W/O5Bi-continuous5O/W. It has also been reported that the permeation profile of the loading drug was highly dependent on the choice and content of the oil phase and the surfactant mixture2. Based on the physicochemical properties of the loading drug, different types of microemulsions can be the optimal carrier. Microemulsion formulation is generally developed with the use of ternary phase diagrams technique2,3. The effect of the formulation parameters is commonly investigated based on the approach which consists in changing one parameter at a time5,6,8. These methods required large number of experiments, a high cost and a long time for the development. Designing drug delivery formulations with the minimum number of trials is very crucial for pharmaceutical scientists10. The fractional factorial design is an experimental design that has been used in formulation of liposomes11,12. This design uses less number of experiments than the full factorial design. In the present study, the fractional factorial design was used to optimize the microemulsion formulation for minoxidil (2,4-diamino-6piperidinopyrimidine 3-oxide; MX). MX was chosen as a model drug for a slightly water soluble compound. It is a potent hypertrichotic agent and widely used for the treatment of androgenic alopecia13. MX is a white to off-white, crystalline powder with molecular weight of 209.25 Da, pKa of 4.6114 and octanol/water partition coefficient (log Ko/w) of 1.2415. For topical use, MX is commercially available in the forms of solution and foam. In the case of topical solution, the mixture solvent of water-propylene glycol (PG)-ethanol (20-20-60% for 2% MX


2

N. Jaipakdee et al.

solution) is used to overcome the poor solubility (Rogaine, Pfizer). In order to enhance the percutaneous absorption of MX, several delivery systems for MX have been developed including the lipid vesicles for example liposomes and ethosomes16,17, niosomes13, as well as nanoparticles18,19. Nevertheless, there is no report regarding the optimization and the permeation characteristics of MX microemulsions yet. In the present study, R(+)-limonene (LMN) was selected as an oil phase because of its ability to increase the skin permeation of a large number of compounds. LMN is a hydrocarbon monoterpene commonly used in transdermal formulations. As for the other terpenes, LMN is classified as generally regarded as safe. The permeation enhancing ability of terpenes was related to the increase of drug diffusivity in and drug partitioning into stratum corneum by disrupting the intercellular lipid bilayers20. Addition of LMN into the oil phase of microemulsions further increased the skin permeation rate21,22. Liu et al.23 reported that LMN microemulsion showed the best transdermal delivery efficacy for curcumin as compared to those of 1,8-cineole and a-terpineol. The curcumin flux for LMN microemulsion was 30and 44-fold higher than those of 1,8-cineole and a-terpineol, respectively. The objective of the present work was to apply the fractional factorial and multi-response optimization designs using desirability function as a fast and efficient approach to develop topical O/ W microemulsion of MX. The effect of various formulation parameters (surfactant ratio, surfactant concentration, co-solvent ratio and oil concentration) on the solubility, permeation flux, lag time and deposition of MX from microemulsions was studied. The level of these formulation factors was optimized in order to obtain the target responses. The physicochemical properties of microemulsions such as pH, viscosity and particle size were also investigated.

Materials and methods Materials Minoxidil (MX) was obtained from S. Tong Chemicals Co., Ltd. (Bangkok, Thailand). R(+)-limonene (LMN) and sodium bis(2ethylhexyl)sulfo-succinate were purchased from FlukaÕ Analytical (St. Louis, MO). Caprylocaproyl polyoxyl-8 glycerides (L.A.SÕ , LAS) was provided by Gattefosse´ SAS (Saint-Priest Cedex, France). Tween 20 (ECOTERIC 20Õ , T20) and Tween 80 (ECOTERIC 80Õ , T80) were purchased from Ajax Finechem Pty Ltd (New South Wales, Australia). PEG-40 hydrogenated castor oil (Nikkol HCO-40Õ , HCO40) was provided by Nikko Chemicals Co., Ltd. (Tokyo, Japan). Ethanol and polyethylene glycol 400 (PEG400) were purchased from Merck (Darmstadt, Germany). Glacial acetic acid and perchloric acid were received from QRe¨C (Auckland, New Zealand). Propylene glycol (PG), acetonitrile and methanol were purchased from RCI Labscan Ltd (Bangkok, Thailand). Deionized water was used throughout the study. All chemicals were used as received. Methods Screening of formulation ingredients Solubility study. Excess amount of MX was added to each surfactant or solvent. The mixture, sealed in light protected and tightly closed test tube, was sonicated for 1 h and then equilibrated at 32 ± 0.5 C in a shaking water bath (Digital Temperature Controller, Polyscience, PA) for at least 24 h until achieved the concentration equilibrium. Mixtures were then filtered through a membrane filter (0.45 lm, 13 mm, Millipore filter, Millipore, MA). The filtrate solution was diluted and analyzed for MX content by HPLC assay as described later.

Pharm Dev Technol, Early Online: 1–12

Surfactant efficiency study. The surfactant efficiency (Smin) which was defined as the minimum amount of surfactant required for completely homogenizing LMN and water was determined24. The test surfactant was added drop by drop to the 1:1 weight ratio of LMN to water mixtures. The amount of surfactant required to change the LMN-water mixture appearance from turbid to transparent, isotropic microemulsion corresponded to the Smin. Experimental design and optimization of MX microemulsion Experimental design. Based on the solubility and surfactant efficacy studies, T20 and LAS were selected as the surfactant system, whereas ethanol and PG were selected as a co-solvent system in the aqueous phase. The effects of the T20 concentration in surfactant system (X1), surfactant system concentration (X2), ethanol concentration in co-solvent system (X3) and LMN concentration (X4) on the solubility of MX (Y1), permeation flux (Y2), lag time (Y3) and deposition (Y4) were studied in a 24-1 fractional factorial design, comprising 8 batch, 24 runs (3 runs for each batch). In order to estimate the experimental error and check the linearity, triplicates were added at the centre point giving a total of 27 runs. The design resolution was IV (four)11, i.e. the main effects are not confounded with two-factor interactions. The levels of the factors are shown in Table 1. The batches were produced in a random order and triplicate measurements were run on each batch to minimize possible systematic errors. Analysis of variance (ANOVA) was used to analyze the main effect of variables on the responses at the p value of 0.05. Moreover, the pareto chart of effects were plotted to show the ANOVA effect estimates plotted against the horizontal axis. This plot will include a vertical line to indicate the p ¼ 0.05 threshold for statistical significance (an effect that exceeds the vertical line may be considered significant)11,25. To establish the correlation between the formulation variables and the responses26, multiple linear regression models were developed using the Minitab 16 Statistical Software (Trial version, Minitab Inc., State College, PA). The data were fitted in Equation (1). Y ¼ 0 þ 1 X1 þ 2 X2 þ 3 X3 þ 4 X4

ð1Þ

where Y stands for the predicted response (solubility of MX, permeation flux, lag time or skin deposition), X1 through X4 stand for the formulation variables (T20 concentration in surfactant system, surfactant system concentration, ethanol concentration in co-solvent system and LMN concentration, respectively), b 1 through b 4 are the respective coefficients and b 0 stands for the intercept or mean. For this design the estimated main effect does Table 1. Fractional factorial design used to optimize MX microemulsion formulations. Level used (%)

Factors (independent variables) X1 T20 concentration in surfactant systema X2 Surfactant system concentration X3 Ethanol concentration in co-solvent systemb X4 LMN concentration

1

1

25 30 25 4

75 50 75 12

Responses (dependent variables) Solubility of MX (mg/mL) Permeation flux (mg/cm2/h) Lag time (h) Deposition (mg/cm2) a

Surfactant system composed of T20 and LAS. Co-solvent system composed of ethanol and PG. Ratio of co-solvent to water was fixed at 1:1 ratio.

b

DOI: 10.3109/10837450.2014.971375

not show the standard errors, because this in a saturated design11,26, where all degrees of freedom is used to estimate the factors main effects and no independent assessment of the error variance is available.

Minoxidil microemulsions using fractional factorial design

3

the intact epidermis was peeled off with forceps, washed with water and kept at 20 C until use (within 1 week)31. The frozen epidermis was thawed at an ambient temperature before use.


Permeation study Multi-response optimization design. A multi-response optimization was performed on the data set of the fractional factorial design using the Minitab 16 Statistical software (Trial version, Minitab Inc., State College, PA). Multi-response optimization technique was considered appropriate for optimizing the control factors. Researchers have advocated the use of desirability function based optimization techniques for multi-response optimization problems27–29. The individual desirability value of the formulation variables were utilized in the present investigation to compute the optimal desirability function for predicting the optimum responses. The optimal desirability function represented the optimal responses. Preparation of microemulsions and MX-loaded microemulsions The tested formulations of O/W liquid microemulsions were prepared by mixing T20 and LAS (surfactant system) with LMN (oil phase). The mixture of T20, LAS and LMN was then added with the required amount co-solvent (ethanol and PG), followed by mixing with water under moderate magnetic stirring for 30 min. For the preparation of MX-loaded microemulsions, required amount of MX (2.0%w/v loading) was added to the prepared microemulsion. The mixture was gently stirred at ambient temperature to obtain a clear, translucent microemulsion. The resulting microemulsions were tightly sealed and stored at ambient temperature. Their physical stability was measured by observing periodically the occurrence of phase separation.

The in vitro permeation of MX from the microemulsions through the pig ear skin was conducted using a side-by-side diffusion cell with a diffusion area of 0.694 cm2 (Crown Glass Company, Somerville, NJ). The system was connected to a water bath maintained at a temperature of 32.0 ± 0.5 C. A thawed skin was mounted between the donor and receptor compartments with a clamp and was hydrated with phosphate buffered saline at pH 7.4 for 1 h. The MX microemulsion (3 mL) was added to the donor compartment, which was in contact with the stratum corneum side of the skin. The receptor compartment was filled with 3 mL of 40% v/v PEG400 solution. At predetermined times, 1.0-mL samples were taken from the receptor compartment and equal volumes of 40% PEG400 solution were immediately added after each sampling. The concentration of MX was analyzed by HPLC. The cumulative amount of drug that permeated the skin was plotted against time. Data analyses The steady state flux (Jss), the permeability coefficient (kp), the concentration gradient (DC) and the drug concentration in the vehicle (Cv) are defined by Equation (2)32: Jss ¼ kp DC ¼ kp Cv

The permeability coefficient from permeation through pig ear skin and the lag time (Tlag) are defined by Equations (3) and (4).

Characterization of MX-loaded microemulsions The physical form and appearance of the tested formulations were investigated visually. Only clear, isotropic one phase systems were considered as microemulsions and further characterized. The pH values of microemulsions were determined at room temperature using a digital pH meter (Corning M250, Ciba Corning Diagnostics Ltd, Suffolk, UK). The viscosity of various microemulsions was measured at 32 C employing a rotating Brookfield viscometer (Model DV-III; Brookfield Engineering Laboratories, Inc., MA) equipped with small sample adapter (SC4-34), at the rotation speed of 30 rpm. The average particle size and polydispersity index were characterized using a Zetasizer Nano (Malvern Instruments, Malvern, Worcestershine, UK) at a temperature of 25 ± 2 C and at 90 to the incident beam applying the principle of photon correlation spectroscopy (PCS). Dispersions were 5-times diluted with prefiltered (0.45 mm) ultrapure water to ensure that the light scattering intensity was within the instrument’s sensitivity range. The solubility of MX in microemulsions was investigated by adding excess amount of MX into 2 mL of each microemulsion. The mixture was sonicated for 1 h and then equilibrated at 32 ± 0.5 C for 24 h using shaking water bath (Digital Temperature Controller, Polyscience, PA). The saturated microemulsion was filtered through nylon syringe filter (0.45 lm, 13 mm, Millipore filter, Millipore, Bedford, MA). The filtrate solution was appropriately diluted and determined by HPLC with reference to a calibration curve. In vitro skin permeation and deposition studies Skin preparation Porcine ears were obtained from a local slaughter house and cleaned with water. After soaking the ears in water at 60 C for 45 s,

ð2Þ

kp ¼

KD h

ð3Þ

h2 6D

ð4Þ

Tlag ¼

where, K is the partition coefficient of drug between the skin and the vehicle; D is the diffusion coefficient of drug in the skin; h is the thickness of the skin. Deposition study After permeation study (10 h), the skin samples were removed and washed on both sides with 0.5 mL of water (3 times) followed by 0.5 mL of methanol (3 times). The part of each skin which directly contacted the MX-loaded microemulsion (area of 0.694 cm2) was separated using scissors and dried at 50 C for 24 h. The dried skin was weighed and cut into small pieces, and extracted for MX with 80%v/v methanol. The extraction was performed twice. The samples were analyzed for MX content by HPLC. HPLC analysis The content of MX dissolved in solvents or surfactants (solubility study), the content of MX permeated through and deposited in the skin (permeation and deposition study) were analyzed by using the method modified from that described in USP30. The HPLC system (Perkin-Elmer, MA) consisted of a UV detector (model 1022 LC plus) and a pump (series 200 LC). The chromatographic separation was achieved on a Hypersil Gold C-18 column (250 4.6 mm, 5 lm; Thermo Electron Corporation, Waltham, MA) with a flow rate of 1 mL/min with UV detection at 254 nm. The mobile phase consisted of methanol: water: glacial acetic acid at a volume ratio of 80:20:1 (pH 3) and 3 g/L of sodium bis(2-ethylhexyl)sulfo-succinate. The retention time of MX was approximately 7 min. The standard curve was linear over a


4

N. Jaipakdee et al.


concentration range of 5.0 to 50.6 lg/mL with an R2 value 40.99. The day-to-day relative standard deviations (RSD) for this assay were less than 5%.

Table 2. Solubility of MX in various surfactants and solvents at 32 C (mean ± SD, n ¼ 3).

Statistical analysis

HCO40 LAS T20 T80 Butanol Ethanol Isopropyl alcohol PG

Surfactants / Solvents

Each experiment was repeated at least three times. The results are expressed as the mean ± S.D. One-way ANOVA was used to test the statistical significance of differences among groups. Statistical significance of the differences of the means was determined by Student’s t-test. All statistical tests were run using the SPSS program for MS Windows, release 19 (SPSS (Thailand) Co. Ltd., Bangkok, Thailand). The significance was determined with 95% confident limits ( ¼ 0.5) and was considered significant at a level of p less than 0.05.

Solubility (mg/mL) 2.89 ± 0.41 12.61 ± 0.22 7.85 ± 0.04 4.00 ± 0.31 16.70 ± 1.04 23.68 ± 1.45 12.22 ± 0.74 125.20 ± 8.81

Table 3. Surfactant efficiency of surfactant (Smin) with 1 to 1 ratio of LMN to water (mean ± SD, n ¼ 3).

Result and discussion Screening of components for ME Drug loading is a critical factor for formulation designation. MX is slightly soluble in water, its water solubility (at 32 C) was 3.0 ± 0.1 mg/mL (mean ± SD, n ¼ 3). The solubility of MX in LMN was also relatively low, less than 0.1 mg/mL. To develop a formulation for a low solubility drug, the components selected should have the ability to solubilize a drug in higher level. Table 2 demonstrates the solubility of MX in various surfactants and solvents. The order of MX solubility in different surfactants occurred in the following: LAS4T204T804HCO40. Only 12.61 ± 0.22 and 7.85 ± 0.04 mg of MX could be dissolved in an mL of LAS and T20, respectively. Therefore, for 2% MX-loaded microemulsion preparations, it was found necessary to add cosolvent into an aqueous phase in order to increase the solubility of the loading drug in the systems. Among the solvent studied, PG and ethanol showed the highest solubility; 125.20 ± 8.81 and 23.68 ± 1.45 mg/mL, respectively. Therefore, both PG and ethanol were selected to use as a co-solvent system in the tested MX microemulsion formulations. In formulation of microemulsions, it is important to determine which surfactants form microemulsions with given oil. The determined amount of LAS required to completely solubilize equal masses of LMN and water was the lowest (45.25 ± 0.52%w/w), followed by HCO40, T20 and T80, respectively (Table 3). Together with the high MX solubilizing capacity, LAS and T20 were therefore selected as a surfactant system for preparation of the tested MX O/W microemulsion formulations. It is known that the hydrophilicity of surfactant influences the type of microemulsions. The low HLB3–6 surfactants are favored for W/O microemulsion, whereas surfactants with high HLB 8–18 are preferred for O/W microemulsion formations3. Both T20 and LAS are non-ionic surfactants with the HLB approximately 16.7 and 14, respectively2,34. They have been used in microemulsions designed for topical delivery because of non-irritant and capability of forming liquid microemulsions with non-alcoholic co-surfactants2. Dependent on the type and concentration of oil phase, the concentration of surfactant used commonly ranged from 18 to 80%2,8. Microemulsion characteristics In order that the effects of formulation component and concentration on the characteristics of microemulsions could be investigated within the limit number of experiments, the 24-1 fractional factorial design was employed. The independent and dependent variables for design-generated experimental runs are given in Table 1. The compositions of the tested formulations are shown in Table 4 and the appearance of the obtained microemulsions is shown in Figure 1. It can be clearly seen that the isotropic,

Surfactants

Smin (%w/w)

HCO40 LAS T20 T80

52.88 ± 1.73 45.25 ± 0.52 54.72 ± 0.99 59.31 ± 0.84

Table 4. Composition of the tested microemulsion formulations. Microemulsion compositions (%w/w) Surfactants Batch code M1 M2 M3 M4 M5 M6 M7 M8 M9

Co-solvent

LMN

a

T20

LAS

Ethanolb

PG

Water

4 12 4 12 8 4 12 12 4

7.5 7.5 12.5 12.5 20 22.5 22.5 37.5 37.5

22.5 22.5 37.5 37.5 20 7.5 7.5 12.5 12.5

8.25 21.75 17.25 4.75 13.0 24.75 7.25 14.25 5.75

24.75 7.25 5.75 14.25 13.0 8.25 21.75 4.75 17.25

33.0 29.0 23.0 19.0 26.0 33.0 29.0 19.0 23.0

a

T20 concentration in surfactant system for batch code M1, M2, M3 and M4 ¼ 25%; batch code M5 ¼ 50%; batch code M6, M7, M8 and M9 ¼ 75%. b Ethanol concentration in co-solvent system for batch code M1, M4, M7 and M9 ¼ 25%; batch code M5 ¼ 50%; batch code M2, M3, M6 and M8 ¼ 75%.

transparent microemulsions could be formed for all formulations except M2 and M7. The cloudy emulsion of M2 and M7 formulations were ascribed to the high LMN content (12%) and insufficient surfactant concentration (30%). Only the transparent microemulsion formulations were loaded with 2%w/v MX and further characterized. The viscosity, pH and particle size of transparent MX microemulsions are shown in Table 5. When compared to the blank microemulsion, the presence of 2% of MX did not affect the characteristics, viscosity and droplet size of microemulsions (data not shown). The viscosity of MX microemulsions ranged from 13.6 ± 1.1 to 55.2 ± 0.9 mPas, a low viscosity indicating an excellent fluidity of microemulsion systems. It was found that the viscosity increased in proportion to T20 concentration in surfactant system, surfactant system and LMN concentrations but decreased in proportion to the ethanol concentration in co-solvent system. The pH of all batches was found to be in the range of 6.8–6.9 which are considered physiologically acceptable. The droplet size was small with all formulations having an average size between 9.0 ± 0.1 to 13.0 ± 0.7 nm. The homogeneity of microemulsion droplets was


DOI: 10.3109/10837450.2014.971375


Figure 1. Appearance of microemulsion formulations.

the

5

tested

Table 5. Viscosity, pH, particle size, MX solubility, permeation flux, lag time and skin deposition of MX after 10 h permeation of the tested MX microemulsions (mean ± SD, n ¼ 3). Batch code

Viscosity (mPas)

pH

Particle size (nm)

Solubility of MX (mg/mL)

Permeation flux (mg/cm2/h)

Lag time (h)

Skin deposition (mg/cm2)

M1 M3 M4 M5 M6 M8 M9

16.4 ± 0.4 17.2 ± 0.5 30.8 ± 0.2 24.5 ± 0.6 13.6 ± 1.1 34.8 ± 2.3 55.2 ± 0.9

6.8 ± 0.1 6.8 ± 0.1 6.9 ± 0.1 6.9 ± 0.1 6.8 ± 0.1 6.9 ± 0.1 6.9 ± 0.1

12.3 ± 0.1 11.6 ± 0.1 13.0 ± 0.7 10.0 ± 0.1 9.0 ± 0.1 9.7 ± 0.2 9.0 ± 0.1

32.9 ± 1.3 31.3 ± 2.9 26.2 ± 0.2 31.1 ± 0.8 36.0 ± 2.0 26.3 ± 0.7 29.5 ± 0.2

85.5 ± 4.4 58.9 ± 6.0 41.9 ± 4.0 44.7 ± 1.4 74.4 ± 8.2 35.4 ± 0.8 29.3 ± 2.9

1.3 ± 0.1 1.3 ± 0.1 1.3 ± 0.2 1.4 ± 0.1 1.4 ± 0.1 1.3 ± 0.0 1.4 ± 0.1

335.0 ± 15.6 254.5 ± 27.5 277.6 ± 24.0 311.4 ± 13.7 315.7 ± 31.1 193.4 ± 18.6 210.1 ± 9.7

confirmed by the low values of polydispersity indexes (less than 0.25) (data not shown). All of the tested microemulsion formulations were physically stable at room temperature in the presence or absence of MX. The changes of isotropic clarity, particle size, and phase separation were not found during 6 months (data not shown). Effect on responses Effect on MX solubility The effects of formulation variables on the solubility of MX in microemulsion were evaluated and shown in Table 5. It was found that the average solubilities of MX were in the range of 26.2 ± 0.2 and 36.0 ± 2.0 mg/mL. The main effect plots and standard pareto chart of all formulation variables for the solubility of MX are shown in Figures 2(a) and 3(a), respectively. As seen in Figure 2(a), the solubility of MX in microemulsion increased in proportion to the concentrations of T20 in surfactant system and ethanol in co-solvent system but decreased in proportion to surfactant system and LMN concentrations. However, the results of the ANOVA and Figure 3(a) show that the influences of only three variables including LMN concentration, surfactant system concentration and concentration of ethanol in co-solvent system are significant (p50.05). The solubility of any compound in the microemulsion generally depends on the solubility of that compound in each ingredient of microemulsion. Considering the solubility of MX in the tested microemulsion components (Table 2), the solubility of MX in LMN (50.1 mg/mL) was much lower than those in the other components. Therefore, increasing LMN concentration caused directly the decrease of MX solubility in the microemulsion. Similar to the LMN concentration, the high amount of surfactant system in the formulation resulted in the lower amount of aqueous phase, and consequently the amount of co-solvent

(ratio of water to co-solvent in the aqueous phase was kept constant at 1:1). The solubility of MX in the surfactant system (T20 and LAS) was lower than that in the co-solvent system (ethanol and PG). Thus, the increasing of surfactant system concentration resulted in the decreasing of MX solubility. Interestingly, even though the solubility of MX in ethanol was 5-times lower than that in PG (Table 2), increasing the concentration of ethanol in co-solvent system led to increase in the solubility of MX in the microemulsion. Figure 4(a) demonstrates the apparent solubility obtained from the experiment and the estimated solubility of MX in the microemulsion. The estimated solubility was the sum of MX solubility in individual component of microemulsion formulations. It can be seen that the batch codes M1, M4 and M9 prepared with 25% of ethanol in co-solvent system showed the comparable values of the apparent and estimated MX solubilities. Increase the concentration of ethanol in co-solvent system resulted in the higher MX solubility of apparent than those of estimated values (Figure 4b). When 75% of ethanol in co-solvent system was used (batch codes M3, M6 and M8), the apparent solubility values were 1.8 times of that of estimated solubility, irrespective of the other component concentrations. This might be attributed to the partitioning and then modifying the solvent characteristics of ethanol. The distribution of alcohols and polyols between the interface and the aqueous phase depends on its hydrophilicity and partition coefficients. The octanol/water partition coefficient as log Po/w of ethanol and PG were 0.31 and 0.92, respectively15. Therefore, it was likely that PG distributed over the aqueous and the interfacial layers, whereas ethanol partitioned over the aqueous, interfacial layer and oil phases35. When the higher concentrations of ethanol (50% and 75% of ethanol in co-solvent system) were used, ethanol would be expected to partition and therefore modify the solvent properties of LMN to be optimally more hydrophilic and of water to be more hydrophobic for MX solubilization.


6

N. Jaipakdee et al.


Figure 2. Main effects plots of T20 concentration in surfactant system (X1), surfactant system concentration (X2), ethanol concentration in co-solvent system (X3) and LMN concentration (X4) for the solubility (a), permeation flux (b), lag time (c) and skin deposition (d) of MX microemulsions.

Effect on permeation flux of MX The effects of formulation variables on the permeation of MX from microemulsion were evaluated and the MX permeation profiles are shown in Figure 5. The permeation flux of all batch codes was shown in Table 5 and was found to be in the range of 29.3 ± 2.9 and 85.5 ± 4.4 mg/cm2/h, indicating the MX permeation characters was significantly affected by the formulation variables. The main effect plots and standard pareto chart of all formulation variables for the permeation fluxes are shown in Figure 2(b) and Figure 3(b), respectively. As can be seen in Figure 2(b), the permeation flux of MX from microemulsion increased in proportion to ethanol concentration in co-solvent system but decreased in proportion to T20 concentration in surfactant system, surfactant system and LMN concentrations. However, the results of the ANOVA and Figure 3(b) showed that the influences of only three variables including surfactant system concentration, concentrations of T20 in surfactant and ethanol in co-solvent systems were significant (p50.05). Numerous reports showed that the permeation fluxes were significantly increased by decreasing the level of surfactant6,8,21,36,37. The skin permeation of ketoprofen increased 12–23 times as the content of surfactant system was decreased from 80 to 30%21. Various mechanisms of this phenomenon have been proposed. The first possible mechanism was attributed to an increased thermodynamic activity of drug in microemulsions at the lower concentration of surfactant and co-surfactant. The thermodynamic activity of drug in the formulation is a significant driving force for the release and permeation of drug into the skin1.

The second mechanism was related to the amount of water content in the formulation that decreased as the surfactant system increased. The relationship between the hydration effect of stratum corneum and the dermal permeation had been reported38. Hydration appears to increase transdermal delivery of both hydrophilic and lipophilic permeants39. Previous reports have demonstrated a poor quercetin permeability from the water-free as compared to those of the water containing microemulsion mixtures, confirming the major role of water in acting as permeation promoter through the membranes40. The third mechanism related to the viscosity of the microemulsion systems. It is known that the viscosity plays an important role in controlling the release of the drug into the receptor. At high surfactant concentration, the progress of emulsification might be compromised by viscous liquid crystalline gel forming at the surfactant– water interface thus leading to the decrease of drug diffusion through the double layer microemulsion to the receptor41. As for the tested MX microemulsions investigated in this study, the increase in permeation fluxes with decreasing surfactant system content, T20 concentration in surfactant system and/or increasing ethanol concentration in co-solvent system did not cause by the increase of thermodynamic activity of MX in microemulsions. On the other hand, the MX solubility and consequently the MX affinity to the system increased as the surfactant content decreased. Similar to the surfactant concentration effect, decreasing the T20 concentration in surfactant system and/or increasing ethanol concentration in co-solvent system individually increased the MX solubility (Figure 2a). Therefore, the thermodynamic activity of MX in the microemulsions was


DOI: 10.3109/10837450.2014.971375


7

Figure 3. Standard pareto charts of the standardized main effect for the solubility (a), permeation flux (b), lag time (c) and skin deposition (d) of MX microemulsions (p ¼ 0.05).

Figure 5. Permeation profiles of the tested MX microemulsion formulations (mean ± SD, n ¼ 3).

Figure 4. The apparent MX solubility (experimental) and the estimated MX solubility in microemulsions (a) and ratio of apparent to estimated MX solubility values as a function of ethanol concentration in cosolvent system (b). The estimated MX solubility were the sum of MX solubility in each component of microemulsion formulations. The good linear relationship between ratio of apparent to estimated MX solubility values and ethanol concentration in cosolvent system was obtained, with the correlation coefficient (R2) of 0.9759.

expected to be lower at low level of surfactant content, T20 concentration in surfactant system and/or at high level of ethanol concentration in co-solvent system. The common effect on MX microemulsions when decreasing the surfactant system content, T20 concentration in surfactant system and/or increasing ethanol concentration in co-solvent system, in addition to enhance the permeation fluxes, was the decrease in the viscosity. Therefore, the possible explanation for the higher permeation flux effect of these variables is related to the lower viscosity and therefore facilitating the rate of drug diffusion in the system. The lower surfactant content resulted in


8

N. Jaipakdee et al.

the lower viscosity as compared to that of the higher content. The lower concentration of T20 also led to decrease the viscosity of microemulsions because of the higher viscosity of T20 as compared to that of LAS (400 and 80110 mPas, respectively)34. The viscosity of ethanol is lower than that of PG (1.1 and 58.1 mPas, respectively)34,42; therefore, a higher concentration of ethanol in co-solvent system resulted in the lower viscosity microemulsions. Additionally, both ethanol and PG can act also as co-surfactants by lowering the interfacial tension of surfactant film2,3. The surface tension of ethanol is lower than that of PG (21.9 and 35.7 mN/m, respectively)42,43. Increasing the ethanol concentration in co-solvent system might lower the interfacial tension of surfactant film in microemulsions, resulting in a more flexible and dynamic layer44,45. The drug in this energy-rich system can diffuse across the flexible and fluidity film. This phenomenon is considered as a thermodynamic process that increases partitioning and diffusion into the stratum corneum37. Considering the permeation parameters in Equation (2), since the MX concentration in all batch codes was fixed, the increase in permeation flux (Jss) was the result of permeability coefficient (kP) increasing. From Equation (3), the permeability coefficient depends on the partition coefficient of drug between skin and formulation (K) and diffusion coefficient of drug in skin (D)32. Therefore, the increase in permeation flux of MX from the microemulsions might be the effect of the increasing of partition coefficient and/or diffusion coefficient. The transdermal delivery is a complex phenomenon governing with the release from the vehicle, the permeation enhancing potency of vehicle and the partitioning into the skin of drug5. Another possibility explaining the higher MX permeation fluxes of those microemulsion variables is related on the mechanism that the microemulsion components are able to enter the skin as monomers, resulting in the increase of drug solubility in the skin1. This process leads to increase the partitioning of drug into the skin creating higher drug concentration gradient, the driving force for drug permeation. The higher MX permeation potential caused by the higher concentration of ethanol in the co-solvent system is probably attributed to the permeation enhancing properties of ethanol itself. Ethanol is a well-known permeation enhancer which reduces the barrier property of skin by a number of mechanisms including lipid fluidization, lipid extraction and effects on lipid ordering as well as effects on keratin33,46. As mention before that decreasing the surfactant system content resulted in the increase of aqueous phase content. Water in aqueous phase could hydrate stratum corneum leading to an increase in drug partition and permeation36. Both T20 and LAS are non-ionic surfactants that can influence the skin barrier, acting as permeation enhancers47. However, the lower viscosity characteristic of LAS might allow the permeation through the skin occurred more easily than T20. MX, which possesses a high affinity for LAS, permeated the skin following this excipient diffusion. Therefore, the lower concentration of T20 in surfactant system (lower ratio of T20 to LAS) increased the microemulsion permeation potential. In the case of LMN content, although the effect was insignificant, a trend of decreasing permeation flux with increasing LMN content was observed. This is attributed to the viscosity of the microemulsions that increased with internal oil content, impeding the drug diffusion in the formulation. Even though, LMN is a permeation enhancer being able to increase the skin permeation of a large number of compounds20; nevertheless, in the case of the tested MX microemulsions, it is considered that high concentration of surfactants and co-solvents that also work as permeation enhancer may make the effect of LMN on the skin permeation less pronounced. Additionally, the solubility of MX in


LMN was much lower than those in co-solvents, surfactants and water. It is possible that most MX accumulated in the continuous phase and at the droplet interfaces, rather than staying in the oil droplet. Therefore, the permeation enhancement by direct drug transfer from microemulsion droplet into stratum corneum may be excluded in the case of MX microemulsions. Effect on lag time The lag time of all batches was shown in Table 5 and was found to be in the narrow average range of 1.3 and 1.4 h. The main effect plots and standard pareto chart of all formulation variables for the lag time of MX are shown in Figure 2(c) and Figure 3(c), respectively. From the results of ANOVA and Figure 3(c), the formulation variables investigated in this study were found to be insignificant effect in the case of lag time. The lag time is defined as the period needed for the permeation rate through the skin to increase until it reaches a stable value. According to Equation (4), the lag time is a permeation parameter that mainly depends on the diffusion coefficient of the drug through the skin5,32. Considering this with no differences between the lag time values obtained after application of different formulations will further indicate that there is no significant difference in the skin diffusion coefficient (D) of all microemulsion formulations. This is in agreement with the study on hydrocortisone microemulsions5. Effect on skin deposition of MX The effects of formulation variables on the skin deposition of MX from microemulsion were evaluated and shown in Table 5. The average skin deposition values of MX were in the range of 193.4 ± 18.6 and 335.0 ± 15.6 mg/cm2. The main effect plots and standard pareto chart of all formulation variables for the skin deposition of MX are shown in Figure 2(d) and Figure 3(d), respectively. As can be seen in Figure 2(d), the skin deposition of MX from microemulsion decreased in proportion to all formulation variables. However, the results of the ANOVA and Figure 3(d) showed that the influences of only two variables including surfactant system concentration and T20 concentration in surfactant system were significant (p50.05). As described in ‘‘Effect on permeation flux of MX’’, the increasing of surfactant system concentration and T20 concentration in surfactant system resulted in either an increase in drug diffusion rate from the formulation and/or drug partitioning between the vehicle and the skin leading to the permeation flux enhancement. However, from ‘‘Effect on lag time’’, the skin diffusivity values of these microemulsion formulations were not increased. These two phenomena resulted in more amount of MX deposited in the skin. Similar results were observed by Butani et al.9 and Liu et al.23. Regression analysis Responses of different batches obtained using half-factorial design are shown in Table 6. Insignificant variables were removed, and adequacy of fitted model was checked by ANOVA. The obtained regression equations are as follows: YSolubility ¼ 30:328 2:039X2 þ 0:845X3 2:079X4

ð5Þ

YPermeation flux ¼ 60:64 9:59X1 19:27X2 þ 5:21X3

ð6Þ

YSkin deposition ¼ 279:61 24:66X1 45:72X2

ð7Þ

The solubility of MX, permeation flux and skin deposition can be calculated by substituting the coded values (1, 0 or +1) of the


DOI: 10.3109/10837450.2014.971375

9

Table 6. Predicted response optimized inputs for target outputs from local and global solutions.


Predicted optimized inputs

Predicted optimized outputs

Local and global solutions

T20a

Surfactant system concentration

Ethanolb

LMN concentration

Solubility of MX

Permeation flux

Skin deposition

Composite desirability

Local Solution Local Solution Local Solution Local Solution Local Solution Local Solution Local Solution Local Solution Local Solution Local Solution Local Solution Local Solution Local Solution Global Solution

25 25 25 50 75 75 75 75 25 25 25 75 75 25

30 30 30 40 30 30 30 30 50 50 50 50 50 30

25 75 25 50 75 25 75 25 75 25 75 75 75 75

4 12 12 8 4 4 12 12 4 4 12 4 12 4

33.600 31.132 29.443 31.089 35.290 33.600 31.132 29.443 31.213 29.523 27.055 31.213 27.055 35.290

84.295 94.708 84.295 44.700 75.533 65.120 75.533 65.120 56.161 45.748 56.161 36.986 36.986 94.708

349.991 349.991 349.991 311.381 300.669 300.669 300.669 300.669 258.552 258.552 258.552 209.230 209.230 349.991

0.99938 0.99917 0.99861 0.97283 0.96639 0.96595 0.96583 0.96520 0.92553 0.92472 0.92408 0.83895 0.83764 0.99975

a

T20 concentration in surfactant system. Ethanol concentration in co-solvent system.

b

variables into Equations (5), (6) and (7). The positive value of the coefficients suggests that the predicted response increases with their associated variables. Whereas the negative value of the coefficients suggests that the predicted response will decrease with the increase in associated variables. The magnitude of the variables indicates the weight of each of these factors. It is observed from Equation (5) that the major governing factors on the solubility of MX are the surfactant system concentration (X2), ethanol concentration in co-solvent system (X3) and LMN concentration (X4) (p50.05). Response surface plots of the solubility of MX are shown in Figure 6(a). At high levels of X2, Y1 decreases from 30.3 to 26.2 mg/mL when X4 increases from 4 to 12% but increases from 27.8 to 28.8 mg/mL when X3 increases from 25% to 75%. From Equation (6), it is observed that the major governing factors on the permeation flux are the T20 concentration in surfactant system (X1), surfactant system concentration (X2) and ethanol concentration in co-solvent system (X3) (p50.05). Response surface plots of the permeation flux of MX are shown in Figure 6(b). At high levels of X2, Y2 decreases from 50.4 to 32.4 mg/cm2/h when X1 increases from 25% to 75% but increases from 35.6 to 47.2 mg/mL when X3 increases from 25% to 75%. It is observed from Equation (7) that the major governing factors on the skin deposition are the T20 concentration in surfactant system (X1) and SAA system concentration (X2) (p50.05). Response surface plot of the skin deposition of MX is shown in Figure 6(c). At high levels of X2, Y4 decreases from 266.0 to 202.0 mg/cm2 when X1 increases from 25% to 75%. Experimental results and predicted values obtained from Equations (5), (6) and (7) are given in Figure 7(a), (b) and (c), respectively. As can be seen, predicted values match experimental values reasonably well with high coefficient of determination for the responses YSolubility, YPermeation flux and YSkin deposition, respectively. Therefore, these models are a sufficient basis for interpretation of the obtained relationships. Determination of the optimum condition for MX microemulsion The optimal desirability function represented the optimal MX solubility, permeation flux and skin deposition of MX microemulsions. In the present study, solubility of MX, permeation flux and skin deposition were jointly optimized using the response optimizer option of MINITAB software26. The ‘‘higher is better’’

criteria was utilized, since the properties of MX microemulsions were considered to be better with high solubility of MX, permeation flux and skin deposition. To accomplish the optimization task, the target responses such as solubility of MX, permeation flux and skin deposition were set with an optimal desirability function having value close to 1. In case of solubility of MX, the target value was set as 37 mg/mL, lower and upper bounds as 26 and 37 mg/mL, respectively. For permeation flux, the target value was set at 100 mg/cm2/h, lower and upper bounds were set at 30 and 100 mg/cm2/h, respectively. Similarly, for skin deposition, the target value was set at 350 mg/cm2, lower and upper bounds were set at 180 and 350 mg/cm2, respectively. The predicted set of input variables with local and global solution in the process of response optimization is given in Table 6. The global solution with composite desirability (D) of 0.99975 exhibited optimized input values with optimized outputs which closely matched with the set target values of output. Characteristics of optimized MX microemulsion The optimized formulation was isotropic, transparent microemulsion with low viscosity. The parameters for physicochemical characters were as follows: 11.5 ± 0.2 nm and 0.14 ± 0.05 for mean droplet size and polydispersity index, 10.4 ± 0.2 cP for viscosity values, 6.7 ± 0.1 for pH value, respectively. Conformity test was conducted using the optimized input values given in the global solution as stated in Table 7 and the results were compared. The result revealed that there is a close agreement between the predicted and achieved optimized values of MX solubility, permeation flux and skin deposition of MX microemulsions.

Conclusion The present study demonstrates the use of a fractional factorial methodology followed by using multi-response optimization technique as a valid method for predicting the solubility, permeation flux and skin deposition in optimization of microemulsion formulations. The derived polynomial equations and the optimal desirability function aid in predicting the values of selected independent variables for preparation of optimal microemulsion formulations with desired properties. By using this approach, MX microemulsion was successfully prepared and optimized. The optimized MX microemulsion formulation having


10

N. Jaipakdee et al.


Figure 6. Effect of variables on the solubility (a), permeation flux (b) and skin deposition (c) of MX microemulsions.

Figure 7. Linear correlation plots between the measured and predicted responses: solubility of MX (a), permeation flux (b) and skin deposition (c).


DOI: 10.3109/10837450.2014.971375

Table 7. Predicted and achieved optimized outputs. Predicted responses

Measured responsesa

% Error

35.29 94.71 349.99

34.65 ± 1.36 97.3 ± 2.5 336.7 ± 12.3

1.8 2.7 3.8

Solubility of MX (mg/mL) Permeation flux (mg/cm2/h) Skin deposition (mg/cm2) a


Expressed as mean ± SD, n ¼ 3.

appropriate physical character with the highest MX solubility, permeation flux and skin deposition was confirmed as low level of LMN content (4%), surfactant content (30%), concentration of T20 in surfactant system (25%) but a high concentration of ethanol in co-solvent system (75%), respectively.

Declaration of interest This study was supported by the Thailand Research Fund (MRG5480035), the Commission of Higher Education and Khon Kaen University, Ministry of Education, Thailand. The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the article.

References 1. Santos P, Watkinson AC, Hadgraft J, Lane ME. Application of microemulsions in dermal and transdermal drug delivery. Skin Pharmacol Physiol 2008;21:246–259. 2. Kreilgaard M. Influence of microemulsions on cutaneous drug delivery. Adv Drug Deliv Rev 2002;54:S77–S98. 3. Lawrence MJ, Rees GD. Microemulsion-based media as novel drug delivery systems. Adv Drug Delivery Rev 2012;64:175–193. 4. Biruss B, Ka¨hlig H, Valenta C. Evaluation of an eucalyptus oil containing topical drug delivery system for selected steroid hormones. Int J Pharm 2007;328:142–151. 5. El Maghraby GM. Transdermal delivery of hydrocortisone from eucalyptus oil microemulsion: effects of cosurfactants. Int J Pharm 2008;355:285–292. 6. Huang YB, Lin YH, Lu TM, Wang RJ, Tsai YH, Wu PC. Transdermal delivery of capsaicin derivative-sodium nonivamide acetate using microemulsions as vehicles. Int J Pharm 2008;349: 206–211. 7. Zhu W, Yu A, Wang W, Dong R, Wu J, Zhai G. Formulation design of microemulsion for dermal delivery of penciclovir. Int J Pharm 2008;360:184–190. 8. Zhang J, Michniak-Kohn B. Investigation of microemulsion microstructures and their relationship to transdermal permeation of model drugs: Ketoprofen, lidocaine and caffeine. Int J Pharm 2011;421: 34–44. 9. Butani D, Yewale C, Misra A. Amphotericin B topical microemulsion: formulation, characterization and evaluation. Colloids Surf B Biointerfaces 2014;116:351–358. 10. Hamed E, Sakr A. Application of multiple response optimization technique to extended release formulations design. J Control Rel 2001;73:329–338. 11. Loukas YL. A computer-based expert system designs and analyzes a 2(kp) fractional factorial design for the formulation optimization of novel multicomponent liposomes. J Pharm Biomed Anal 1998;17: 133–140. 12. Stensrud G, Sande SA, Kristensen S, Smistad G. Formulation and characterization of primaquine loaded liposomes prepared by a pH gradient using experimental design. Int J Pharm 2000;198:213–228. 13. Balakrishnan P, Shanmugam S, Lee WS, et al. Formulation and in vitro assessment of minoxidil niosomes for enhanced skin delivery. Int J Pharm 2009;377:1–8. 14. Gibson G, Ramstad T, Mills KA, Dunn MJ. A method for the determination of minoxidil in hair-regrowth formulations by micellar electrokinetic capillary chromatography. Farmaco 2005;60: 847–53.

11

15. Hansch C, Leo A, Hoekman D. Exploring QSAR: hydrophobic, electronic and steric constants. Washington, DC: American Chemical Society, 1995. 16. Mura S, Manconi M, Valenti D, et al. Transcutol containing vesicles for topical delivery of minoxidil. J Drug Target 2011;19: 189–96. 17. Lo´pez-Pinto JM, Gonza´lez-Rodrı´guez ML, Rabasco AM. Effect of cholesterol and ethanol on dermal delivery from DPPC liposomes. Int J Pharm 2005;298:1–12. 18. Kwon TK, Kim JC. In vitro skin permeation of monoolein nanoparticles containing hydroxypropyl b-cyclodextrin/minoxidil complex. Int J Pharm 2010;392:268–273. 19. Padois K, Cantie´ni C, Bertholle V, et al. Solid lipid nanoparticles suspension versus commercial solutions for dermal delivery of minoxidil. Int J Pharm 2011;416:300–304. 20. Sapra B, Jain S, Tiwary AK. Percutaneous permeation enhancement by terpenes: mechanistic view. AAPS J 2008;10: 120–132. 21. Rhee YS, Choi JG, Park ES, Chi SC. Transdermal delivery of ketoprofen using microemulsions. Int J Pharm 2001;228: 161–170. 22. Ngawhirunpat T, Worachun N, Opanasopit P, et al. Cremophor RH40-PEG 400 microemulsions as transdermal drug delivery carrier for ketoprofen. Pharm Dev Technol 2013;18:798–803. 23. Liu C-H, Chang F-Y, Hung D-K. Terpene microemulsions for transdermal curcumin delivery: effects of terpenes and cosurfactants. Colloids Surf B Biointerf 2011;82:63–70. 24. Djekic L, Primorac M. The influence of cosurfactants and oils on the formation of pharmaceutical microemulsions based on PEG-8 caprylic/capric glycerides. Int J Pharm 2008;352:231–239. 25. Zhang D, Wang R, Yang X. Application of fractional factorial design of ZSM-5 synthesis using ethanol as template. Micropor Mesopor Mater 2009;126:8–13. 26. Ryan B, Joiner B, Cryer J. MINITAB handbook: updated for release 14, 5th ed. California: Thomson Learning; 2005. 27. Candioti LV, De Zan MM, Ca´mara MS, Goicoechea HC. Experimental design and multiple response optimization. Using the desirability function in analytical methods development. Talanta 2014;124:123–138. 28. Sarkar S, Sekh M, Bhattacharyya B. Modeling and optimization of wire electrical discharge machining of g–TiAl in trim cutting operation. J Mater Process Technol 2008;205:376–387. 29. Yigezu BS, Mahapatra MM, Jha PK. On modeling the abrasive wear characteristics of in situ Al-12%Si/TiC composites. Mater Design 2013;50:277–284. 30. The United States pharmacopeial convention. The United States pharmacopeia, USP 29/The National Formulary, NF24. Rockville: The United States Pharmacopeial Convention; 2005. 31. Bhatia KS, Gao S, Singh J. Effect of penetration enhancers and iontophoresis on the FT-IR spectroscopy and LHRH permeability through porcine skin. J Contr Rel 1997;47:81–89. 32. Mitragotri S, Anissimov YG, Bunge AL, et al. Mathematical model of skin permeability: an overview. Int J Pharm 2011;418: 115–129. 33. Lane ME. Skin penetration enhancers. Int J Pharm 2013;447: 12–21. 34. Rowe RC, Sheskey PJ, Quinn ME. Handbook of pharmaceutical excipients, 6th ed. Illinois: Pharmaceutical Press; 2009. 35. Chaiyana W, Saeio K, Hennink WE, Okonogi S. Characterization of potent anticholinesterase plant oil based microemulsion. Int J Pharm 2010;401:32–40. 36. Patel MR, Patel RB, Parikh JR, et al. Effect of formulation components on the in vitro permeation of microemulsion drug delivery system of fluconazole. AAPS PharmSciTech 2009;10: 917–923. 37. Tsai Y-H, Lee K-F, Huang Y-B, et al. In vitro permeation and in vivo whitening effect of topical hesperetin microemulsion delivery system. Int J Pharm 2010;388:257–262. 38. Mohammed C, Manoj V. Aerosol-OT microemulsions as transdermal carriers of tetracaine hydrochloride. Drug Dev Ind Pharm 2000; 26:507–572. 39. Williams AC, Barry BW. Penetration enhancers. Adv Drug Deliv Rev 2004;56:603–618. 40. Censi R, Martena V, Hoti E, et al. Permeation and skin retention of quercetin from microemulsions containing TranscutolÕ P. Drug Dev Ind Pharm 2012;38:1128–1133.

12

N. Jaipakdee et al.


41. Zidan AS, Sammour OA, Hammad MA, et al. Quality by design: understanding the formulation variables of a cyclosporine A self-nanoemulsified drug delivery systems by Box-Behnken design and desirability function. Int J Pharm 2007;332:55–63. 42. Khattab IS, Bandarkar F, Fakhree MAA, Jouyban A. Density, viscosity, and surface tension of water+ethanol mixtures from 293 to 323 K. Korean J Chem Eng 2012;29:812–817. 43. Hoke Jr BC, Patton EF. Surface tensions of propylene glycol + water. J Chem Eng Data 1992;37:331–333. 44. Trotta M, Gallarate M, Pattarino F, Carlotti ME. Investigation of the phase behaviour of systems containing lecithin


and 2-acyl lysolecithin derivatives. Int J Pharm 1999;190: 83–89. 45. Wu H, Ramachandran C, Weiner ND, Roessler BJ. Topical transport of hydrophilic compounds using water-in-oil nano emulsions. Int J Pharm 2001;220:63–75. 46. Kadir R, Stempler D, Liron Z, Cochen S. Delivery of theophylline into excised human skin from alkanoic acid solution: a ‘‘push-pull’’ mechanism. J Pharm Sci 1987;76:774–779. 47. Nokhodchi N, Shokri J, Dashbolaghi A, et al. The enhancement effect of surfactants on the penetration of lorazepam through rat skin. Int J Pharm 2003;250:359–369.

Optimization of thiamethoxam adsorption parameters using multi-walled carbon nanotubes by means of fractional factorial design.

A quality by design approach to optimization of emulsions for electrospinning using factorial and D-optimal designs.

Mucoadhesive hydrogel films of econazole nitrate: formulation and optimization using factorial design.

Amphotericin B-loaded polymeric nanoparticles: formulation optimization by factorial design.

Real-time PCR probe optimization using design of experiments approach.

Novel Starch-PVA Polymer for Microparticle Preparation and Optimization Using Factorial Design Study.

Design and optimization of bilayered tablet of Hydrochlorothiazide using the Quality-by-Design approach.

Factorial Design Approach in Proportioning Prestressed Self-Compacting Concrete.

Development and optimization of enteric coated mucoadhesive microspheres of duloxetine hydrochloride using 3(2) full factorial design.

Preparation, optimization and characterization of bovine lactoferrin-loaded liposomes and solid lipid particles modified by hydrophilic polymers using factorial design.

Factorial design applied to the optimization of lipid composition of topical antiherpetic nanoemulsions containing isoflavone genistein.

Nanoemulsions containing a synthetic chalcone as an alternative for treating cutaneous leshmaniasis: optimization using a full factorial design.

Optimization of HPMC and carbopol concentrations in non-effervescent floating tablet through factorial design.

Full factorial design for optimization, development and validation of HPLC method to determine valsartan in nanoparticles.

Optimization of Chitosan Microspheres Spray Drying via 32 Full Factorial Design.

Optimal design of one-dimensional photonic crystal filters using minimax optimization approach.

Optimization of the ethanol recycling reflux extraction process for saponins using a design space approach.

Process-time Optimization of Vacuum Degassing Using a Genetic Alloy Design Approach.

A Fractional Factorial Design to Study the Effect of Process Variables on the Preparation of Hyaluronidase Loaded PLGA Nanoparticles.

Using blocked fractional factorial designs to construct discrete choice experiments for healthcare studies.

Fixed structure compensator design using a constrained hybrid evolutionary optimization approach.

Fractional Factorial Design Study on the Performance of GAC-Enhanced Electrocoagulation Process Involved in Color Removal from Dye Solutions.

Nanostructured lipid carriers as a potential vehicle for Carvedilol delivery: Application of factorial design approach.

What is a factorial study design?